Power utilization change analysis method and device based on rhythm decomposition and medium

Abstract

The invention relates to an electricity utilization change analysis method and device based on rhythm decomposition and a medium, and the method comprises the following steps: obtaining original time series data of electric power and electric quantity, and carrying out the preprocessing to obtain processed data; based on a predefined multiplication decomposition mathematical model, decomposing the processed data under different rhythm elements to obtain a trend component of electric power and electric quantity, a fluctuation component under different rhythm elements and a sudden random non-rhythm component; analyzing the trend component, the fluctuation component and the random non-rhythm component to obtain each refined component; combining the refined components based on a multiplication decomposition model, training the model by using the preprocessed data set, and constructing a final power consumption analysis model; and analyzing the new data by using the final power utilization analysis model to obtain a power utilization analysis result. Compared with the prior art, the method has the advantages of high accuracy and the like.

Description

Technical Field

[0001] The present invention relates to the field of energy technology, and in particular to a method, device and medium for analyzing electricity consumption changes based on rhythm decomposition. Background Art

[0002] Since electricity cannot be stored in large quantities, its generation, transmission, distribution and use must be carried out simultaneously to ensure that the output of electricity follows the actual demand curve and maintain a dynamic balance between supply and demand. Therefore, for power companies, the timing analysis of electricity and electricity plays a key role in formulating economic, reliable and safe power system operation strategies, and has therefore become an increasingly important focus.

[0003] Many existing technologies for time series analysis and prediction of power data mainly rely on black box models such as neural networks, which have certain limitations. In view of the fact that power demand is affected by a variety of complex factors, it is necessary to describe these factors through methods with strong interpretability. Furthermore, in the process of deconstructing power time series, traditional methods often only rely on "correlation" determinism, and rarely conduct in-depth analysis of power demand from the perspective of rhythm to extract relevant factors.

[0004] To sum up, it is still necessary to analyze the rhythmic characteristics of power and electricity data, complete the rhythmic decomposition of power and electricity data, improve the interpretability of the model, and accurately characterize different explicit and implicit electricity consumption elements. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a method, device and medium for analyzing electricity consumption changes based on rhythm decomposition.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] A method for analyzing electricity consumption changes based on rhythm decomposition includes the following steps:

[0008] S1: Obtain the original time series data of electric power and perform preprocessing to obtain processed data;

[0009] S2: Based on the predefined multiplication decomposition mathematical model, the processed data is decomposed under different rhythmic elements to obtain the trend component of power, the fluctuation component under different rhythmic elements, and the sudden random non-rhythmic component;

[0010] S3: Analyze the trend component, fluctuation component and random non-rhythmic component to obtain the refined components;

[0011] S4: The refined components are merged based on the multiplication decomposition model, and the model is trained using the preprocessed data set to construct the final power consumption analysis model;

[0012] S5: Use the final electricity consumption analysis model to analyze the new data and obtain the electricity consumption analysis results.

[0013] Furthermore, the original time series data includes the granularity of the time series, the maximum temperature, the minimum temperature, the power quantity, the average temperature and the random non-rhythmic component.

[0014] Furthermore, preprocessing includes data access, granularity calibration and exception handling.

[0015] Furthermore, the particle size calibration is specifically as follows:

[0016] The original time series data is calibrated by piecewise linear interpolation, and the calculation expression is:

[0017]

[0018] Where (t0, y0) and (t1, y1) are the original time series data to be calibrated, t is the target time series granularity, and y is the interpolation result at the target time series granularity.

[0019] Furthermore, exception handling includes the following cases:

[0020] Through type detection, missing power data and non-numerical data are discovered and filled; for power data on non-weekends and non-holidays, sliding average is used for filling; for power data on weekends or holidays, similar days are used for filling;

[0021] Through local outlier detection, the power data is judged to be abnormal and replaced with the average value of the adjacent days;

[0022] Through zero-value detection, zero-value temperature data is found and filled with historical meteorological data;

[0023] Through nearest neighbor comparison, the temperature data is judged to be abnormal and filled with historical meteorological data.

[0024] Furthermore, the predefined multiplication decomposition mathematical model is specifically:

[0025] Original time series data = trend component × (1 + volatility component + random component)

[0026] Wherein, the trend component is the rhythmic trend related to the power data and the development rhythm of the region, the fluctuation component is the rhythmic change of power over time, and the random component is the random non-rhythmic component.

[0027] The expression of the trend component is:

[0028] F(X1(t)…,X i (t),…)=∑ i β i X i (t)+b

[0029] In the formula, F(X1(t)…,X i (t),…) is the trend component, β i represents the weight corresponding to each factor, and b is the bias;

[0030] The expression of the fluctuation component is:

[0031] Fluctuation component = y(t) + w(t) + α T T(t)

[0032] Where y(t) is the annual rhythm fluctuation, w(t) is the weekly rhythm fluctuation, and T(t) is the rhythm fluctuation caused by temperature;

[0033] The expression of the random component is:

[0034] Random component = h(t)+∈ t

[0035]

[0036] Z(t)=[1(t∈D1),…1(t∈D L )]

[0037] Where h(t) represents the impact of random non-rhythmic components, L is the number of predefined emergency event intervals, and κ l is a trainable parameter, ∈ t is the residual component representing the error.

[0038] Furthermore, the power analysis model is used to evaluate the model accuracy through the root mean square error, and the calculation expression is:

[0039]

[0040] Where RMSE is the root mean square error; i is the actual value, is the predicted value.

[0041] Furthermore, the power consumption analysis model evaluates the accuracy of the model through the mean absolute error, and the calculation expression is:

[0042]

[0043] Where, MAE is the mean absolute error; yi is the actual value, is the predicted value.

[0044] A second aspect of the present invention is a power consumption analysis device based on rhythmic power decomposition, comprising a memory, a processor, and a program stored in the memory, wherein when the processor executes the program, it implements any of the above methods for analyzing power consumption changes based on rhythmic decomposition.

[0045] A third aspect of the present invention is a storage medium having a program stored thereon, which, when executed, implements any of the above methods for analyzing electricity consumption changes based on rhythm decomposition.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] 1) The present invention is based on the rhythmic characteristics of electric power. By multiplying and decomposing the trend component, the fluctuation component and the random non-rhythmic component, each component is established on the basic trend change of electric power. By calculating through the power consumption analysis model, it can more accurately describe the coupling relationship between different components, make the law of power consumption change clearer and more accurate, and more accurately describe the relationship between the change of electric power and the rhythmic properties, which is helpful to analyze and describe the law of power consumption.

[0048] 2) The present invention can extract and summarize the periods, amplitudes, phases, trends, and non-rhythmic features of different components that cannot be described by the above characteristics, and characterize the electricity demand patterns of the overall region and even a certain industry under it, so as to better meet the needs of power system planning and operation and provide reliable technical support. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic diagram of the process of the present invention.

[0050] Figure 2 It is a schematic diagram of the multiplication decomposition result of the present invention.

[0051] Figure 3 It is a schematic diagram of the refined components of the present invention. DETAILED DESCRIPTION

[0052] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0053] Example 1

[0054] The present invention is a method, device and medium for analyzing power consumption changes based on rhythm decomposition, such as Figure 1 As shown, the method comprises the following steps:

[0055] S1: Obtain the original time series data of electric power and perform preprocessing to obtain processed data;

[0056] Preprocessing includes three parts: data access, granularity calibration and exception handling, specifically:

[0057] Data access: In this embodiment, the time series data used needs to contain the following features:

[0058] 1) Granularity of time series: Taking "day" as an example, each observation in the original time series data describes the characteristics of temperature, power consumption, etc. on a certain day;

[0059] 2) Maximum temperature: one of the characteristics of the observations in the sequence, representing the highest temperature of the day;

[0060] 3) Minimum temperature: One of the characteristics of the observations in the sequence, representing the lowest temperature of the day;

[0061] 4) Electricity / power: One of the characteristics of the observations in the sequence, representing the total peak power load / total power consumption in the area on that day;

[0062] 5) Average temperature: One of the characteristics of the observations in the sequence, representing the average temperature of the day.

[0063] 6) Other statistical indicators of social and economic levels, such as GDP, industrial added value, total import and export trade, employment rate, greenhouse gas emissions, etc.

[0064] Granularity calibration: In the process of data access, the granularity of all time series objectively differs to a certain extent. For example, it is difficult to obtain the original data of GDP with daily granularity. Therefore, it is necessary to calibrate these data towards the target granularity. The method is piecewise linear interpolation, and the expression is:

[0065]

[0066] Where (t0, y0) and (t1, y1) are the original time series data to be calibrated, t is the target time series granularity, and y is the interpolation result at the target time series granularity.

[0067] Exception handling includes the following steps:

[0068] 1) Through type detection, missing power data and non-numerical data are found and filled; for non-weekend and non-emergency power data, sliding average is used for filling; for weekend or emergency power data, similar days are used for filling;

[0069] 2) Through local outlier detection, the power data is judged to be abnormal and replaced with the average value of the adjacent day;

[0070] 3) Through zero value detection, zero value temperature data is found and filled with historical meteorological data;

[0071] 4) Determine temperature data anomalies through nearest neighbor comparison and fill them with historical meteorological data.

[0072] S2: Based on the predefined multiplication decomposition mathematical model, the processed data is decomposed under different rhythmic elements to obtain the trend component of power, the fluctuation component under different rhythmic elements, and the sudden random non-rhythmic component;

[0073] The predefined mathematical model of multiplication decomposition is as follows:

[0074] Original time series data = trend component × (1 + volatility component + random component)

[0075] Wherein, the trend component is the rhythmic trend related to the power data and the development rhythm of the region, the fluctuation component is the rhythmic change of power over time, and the random component is the random non-rhythmic component.

[0076] The multiplicative decomposition algorithm is used to model the time series of electric power. The trend component reflects the rhythmic trend related to the development rhythm of the power data and the region, the fluctuation component represents the rhythmic changes of power over time, and the random component describes the impact of non-rhythmic factors on power that cannot be described by explicit laws.

[0077] The expression of the trend component is:

[0078] F(X1(t)…,X i (t),…)=∑ i β i X i (t)+b

[0079] In the formula, F(X1(t)…,X i (t),…) is the trend component, β i represents the weight corresponding to each factor, and b is the bias;

[0080] The expression of the fluctuation component is:

[0081] Fluctuation component = y(t) + w(t) + α T T(t)

[0082] Where y(t) is the annual rhythm fluctuation, w(t) is the weekly rhythm fluctuation, and T(t) is the rhythm fluctuation caused by temperature. For T(t), the preprocessed meteorological data can be directly used for fitting, while for y(t) and w(t), Fourier series can be used:

[0083]

[0084] Modeling is performed, where P is the period of the rhythm, α n , α N+n and N are both trainable parameters.

[0085] The expression of the random component is:

[0086] Random component = h(t)+∈ t

[0087]

[0088] Z(t)=[1(t∈D1),…1(t∈D L )]

[0089] Where h(t) represents the impact of random non-rhythmic components, L is the number of predefined emergency event intervals, and κ l is a trainable parameter, ∈ t It is the residual component representing the error. By calculating its autocorrelation coefficient or using white noise test algorithms such as power spectral density analysis, the randomness of the random component can be trusted.

[0090] S3: Analyze the trend component, fluctuation component and random non-rhythmic component to obtain the refined components;

[0091] like Figure 2 As shown, it is the final effect diagram of each component after refinement. It can be seen that after the original power and electricity time series data is decomposed by the method described in this patent, the final result of each rhythmic component. It can be seen that the overall growth trend of the "trend component" is consistent with the economic construction and social development trend of the city's infrastructure; the "annual rhythm component" can intuitively see that in the same rhythmic cycle, the electricity demand in the high temperature and low temperature sections has surged, while the electricity demand in the period with milder temperature is at the bottom; the "weekly rhythm component" can intuitively reflect the significant difference in electricity demand on weekdays compared to weekends in the same rhythmic cycle: the "residual component" conforms to the random characteristics of "white noise"; in particular, through the historical trend of the emergency component, the impact of the two periods (early 20 and early 22) on power and electricity can be clearly seen.

[0092] like Figure 3As shown, the refined components include the impact of sudden events, white noise error residual components, annual rhythm fluctuations, weekly rhythm fluctuations, temperature rhythm fluctuations and socioeconomic indicators.

[0093] S4: The refined components are merged based on the multiplication decomposition model, and the model is trained using the preprocessed data set to construct the final power consumption analysis model;

[0094] The method for selecting and evaluating the power consumption analysis model can be the root mean square error or the mean absolute error. The calculation expression of the root mean square error is:

[0095]

[0096] Where RMSE is the root mean square error; i is the actual value, is the predicted value. When the actual value is completely consistent with the predicted value, it is equal to 0, which is a perfect model. The larger the error, the larger the value.

[0097] The calculation expression of mean absolute error is:

[0098]

[0099] Where, MAE is the mean absolute error; y i is the actual value, is the predicted value. When the actual value is completely consistent with the predicted value, it is equal to 0, which is a perfect model. The larger the error, the larger the value.

[0100] S5: Use the final electricity consumption analysis model to analyze the new data and obtain the electricity consumption analysis results.

[0101] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0102] In the context of the present invention, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0103] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A method for analyzing electricity consumption changes based on rhythm decomposition, characterized in that: The following steps are involved: S1: Obtain the original time series data of electric power and perform preprocessing to obtain processed data; S2: Based on the predefined multiplication decomposition mathematical model, the processed data is decomposed under different rhythmic elements to obtain the trend component of power, the fluctuation component under different rhythmic elements, and the sudden random non-rhythmic component; S3: Analyze the trend component, fluctuation component and random non-rhythmic component to obtain the refined components; S4: The refined components are merged based on the multiplication decomposition model, and the model is trained using the preprocessed data set to construct the final power consumption analysis model; S5: Use the final electricity consumption analysis model to analyze the new data and obtain the electricity consumption analysis results.

2. The method for analyzing electricity consumption changes based on rhythm decomposition according to claim 1, characterized in that: The original time series data includes the granularity of the time series, the maximum temperature, the minimum temperature, the power quantity, the average temperature and the random non-rhythmic component.

3. The method for analyzing electricity consumption changes based on rhythm decomposition according to claim 1, characterized in that: The preprocessing includes data access, granularity calibration and exception processing.

4. The method for analyzing electricity consumption changes based on rhythm decomposition according to claim 3 is characterized in that: The particle size calibration is specifically: The original time series data is calibrated by piecewise linear interpolation, and the calculation expression is: Where (t0, y0) and (t1, y1) are the original time series data to be calibrated, t is the target time series granularity, and y is the interpolation result at the target time series granularity.

5. The method for analyzing electricity consumption changes based on rhythm decomposition according to claim 3 is characterized in that: The exception handling includes the following situations: Through type detection, missing power data and non-numerical data are discovered and filled; for power data on non-weekends and non-holidays, sliding average is used for filling; for power data on weekends or holidays, similar days are used for filling; Through local outlier detection, the power data is judged to be abnormal and replaced with the average value of the adjacent days; Through zero-value detection, zero-value temperature data is found and filled with historical meteorological data; Through nearest neighbor comparison, the temperature data is judged to be abnormal and filled with historical meteorological data.

6. The method for analyzing electricity consumption changes based on rhythm decomposition according to claim 1, characterized in that: The predefined multiplication decomposition mathematical model is specifically: Original time series data = trend component × (1 + volatility component + random component) Wherein, the trend component is the rhythmic trend related to the power data and the development rhythm of the region, the fluctuation component is the rhythmic change of power over time, and the random component is the random non-rhythmic component. The expression of the trend component is: F(X1(t)…,X i (t),…)=Σ i β i X i (t)+b In the formula, F(X1(t)…,X i (t),…) is the trend component, β i represents the weight corresponding to each factor, and b is the bias; The expression of the fluctuation component is: Fluctuation component = y(t) + w(t) + α T T(t) Where y(t) is the annual rhythm fluctuation, w(t) is the weekly rhythm fluctuation, and T(t) is the rhythm fluctuation caused by temperature; The expression of the random component is: Random component = h(t)+∈ t Z(t)=[1(t∈D1),…1(t∈D L )] Where h(t) represents the impact of random non-rhythmic components, L is the number of predefined emergency event intervals, and k l is a trainable parameter, ∈ t is the residual component representing the error.

7. The method for analyzing electricity consumption changes based on rhythm decomposition according to claim 1, characterized in that: The power consumption analysis model evaluates the model accuracy through the root mean square error, and the calculation expression is: Where RMSE is the root mean square error; i is the actual value, is the predicted value.

8. The method for analyzing electricity consumption changes based on rhythm decomposition according to claim 1, characterized in that: The power consumption analysis model evaluates the model accuracy through the mean absolute error, and the calculation expression is: Where, MAE is the mean absolute error; y i is the actual value, is the predicted value.

9. An electricity consumption analysis device based on rhythmic power decomposition, comprising a memory, a processor, and a program stored in the memory, characterized in that: When the processor executes the program, it implements a method for analyzing electricity consumption changes based on rhythm decomposition as described in any one of claims 1-8.

10. A storage medium having a program stored thereon, characterized in that: When the program is executed, it implements a method for analyzing electricity consumption changes based on rhythm decomposition as described in any one of claims 1-8.

Images